The first electrical energy distribution systems, designed over a century ago, bore the hallmarks of centralized generation and unilateral flow of power. Among the problems associated with early power distribution systems were the hazards of direct current electricity, the isolated nature of each distribution network, the difficulty in predicting demand, the potential for cascading failures caused by discrete breakdowns, and the inefficient transmission of power over long distances.
Some of these problems have been successfully addressed in later years. For example, the switch to alternating current electricity made long-distance power transmission more safe and efficient, and new power grid topologies made distribution less vulnerable to catastrophic failure. Yet, some of these same problems currently remain in the industry.
In particular, measuring, predicting, and planning for changes in consumer demand for energy has proven difficult since the beginning, and remains so today. Despite this difficulty, accurate determinations of demand are important for utility companies, because imbalances between production and consumption can cause brownouts, and even blackouts. Given the short timeframe between production and consumption, these imbalances can become disastrous almost as soon as they are detected, if not before detection. Utility companies are also forced to maintain the generation and distribution capacity to satisfy peak loads, even if such loads occur very infrequently. A delicate balance, therefore, must be struck between overproduction and underproduction lest grid failures occur.